Xia 

et al. BMC Public Health (2023) 23:149 BMC Public Health

https://doi.org/10.1186/s12889-022-14931-x

RESEARCH Open Access

Effects of ambient temperature on mortality

among elderly residents of Chengdu
city in Southwest China, 2016–2020:
a distributed‑lag non‑linear time series analysis
Yizhang Xia1,2†, Chunli Shi1†, Yang Li1, Xianyan Jiang1, Shijuan Ruan1, Xufang Gao3, Yu Chen2, Wei Huang4,
Mingjiang Li5, Rong Xue6, Xianying Wen7, Xiaojuan Peng8, Jianyu Chen1* and Li Zhang1* 

Abstract 
Background  With complex changes in the global climate, it is critical to understand how ambient temperature
affects health, especially in China. We aimed to assess the effects of temperature on daily mortality, including total
non-accidental, cardiovascular disease (CVD), respiratory disease, cerebrovascular disease, and ischemic heart disease
(IHD) mortality between 2016 and 2020 in Chengdu, China.
Methods  We obtained daily temperature and mortality data for the period 2016–2020. A Poisson regression model
combined with a distributed-lag nonlinear model was used to examine the association between temperature and
daily mortality. We investigated the effects of individual characteristics by sex, age, education level, and marital status.
Results  We found significant non-linear effects of temperature on total non-accidental, CVD, respiratory, cerebro-
vascular, and IHD mortality. Heat effects were immediate and lasted for 0–3 days, whereas cold effects persisted for
7–10 days. The relative risks associated with extreme high temperatures (99th percentile of temperature, 28 °C) over
lags of 0–3 days were 1.22 (95% confidence interval [CI]: 1.17, 1.28) for total non-accidental mortality, 1.40 (95% CI:
1.30, 1.50) for CVD morality, 1.34 (95% CI: 1.24, 1.46) for respiratory morality, 1.33 (95% CI: 1.20, 1.47) for cerebrovascular
mortality, and 1.38 (95% CI: 1.20, 1.58) for IHD mortality. The relative risks associated with extreme cold temperature
(1st percentile of temperature, 3.0 °C) over lags of 0–14 days were 1.32 (95% CI: 1.19, 1.46) for total mortality, 1.45 (95%
CI: 1.24, 1.68) for CVD morality, 1.28 (95% CI: 1.09, 1.50) for respiratory morality, 1.36 (95% CI: 1.09, 1.70) for cerebrovas-
cular mortality, and 1.26 (95% CI: 0.95, 1.68) for IHD morality. We found that hot and cold affects were greater in those
over 85 years of age, and that women, individuals with low education levels, and those who were widowed, divorced,
or never married, were more vulnerable.
Conclusions  This study showed that exposure to hot and cold temperatures in Chengdu was associated with
increased mortality, with people over 85 years old, women, those with low education levels, and unmarried individu-
als being more affected by hot and cold temperatures.

†
Yizhang Xia and Chunli Shi made the same contributions to this manuscript.
*Correspondence:
Jianyu Chen
aculacjy@163.net
Li Zhang
657096242@qq.com
Full list of author information is available at the end of the article

© The Author(s) 2023. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​
mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Xia et al. BMC Public Health (2023) 23:149
Keywords  Temperature, Mortality, Distributed-lag nonlinear model
Background
Along with climate change, ambient temperature is a
major environmental health risk factor that is consid-
ered to be a major public health concern [1–3]. The
International Panel on Climate Change predicts that
global climate change will result in an increased risk of
temperature-related mortality worldwide [4]. Therefore,
improved understanding of the relationship between
temperature and human health may provide important
insights to address the health hazards caused by climate
change.
Many studies have shown that high and low tempera-
tures have significant effects on human health [5–8]. For
instance, a study in Beijing showed that an increase in the
6-day moving average temperature from moderately hot
(30.2 °C) to extremely hot (36.9 °C) resulted in a signifi-
cant increase in cardiovascular disease (CVD) admissions
of 16.1% (95% confidence interval [CI]: 12.8%–28.9%)
[9]. Another study in Thessaloniki, Greece reported
risk findings for cardiovascular mortality (increase in
risk per degree increase in temperature: 4.4%, 95% CI:
2.7%–6.1%) and respiratory mortality (increase in risk
per degree increase in temperature: 5.9%, 95% CI: 1.8%–
10.3%) at lags of 0–1  days [2]. Additionally, studies sug-
gest that people aged over 65 years are more vulnerable
to the health risks of climate change [10].
Current studies regarding the effect of temperature
on mortality are mainly from developed countries. As
the world’s largest developing country and second larg-
est economy, China’s rapid economic development has
brought with it a series of environmental health prob-
lems. However, studies regarding the effect of tempera-
ture on mortality in China have only been conducted
in a few eastern regions, such as Beijing [11], Tianjin
[1], Shanghai [12], and Shenzhen [13], with few studies
conducted in western regions, especially in Chengdu.
In particular, the effect of low and high temperatures
on mortality of people aged over 65  years in Chengdu
remains unclear. Therefore, further studies would be
urgently needed for a comprehensive understanding of
the adverse effects of extreme cold and heat.
In this study, we aimed to determine the impact of
extreme temperature on mortality, including total non-
accidental, cardiovascular, respiratory, cerebrovascular,
and ischemic heart disease (IHD) mortality, in Chengdu,
China. We also examined the vulnerable populations by
age, gender, education level, and marital status in a strati-
fied analysis. Our findings could provide a new theoreti-
cal basis for climate-related public health policies.
Page 2 of 12
Methods
Study area and population
Chengdu is located in the western portion of the Sichuan
Basin and the upstream Yangzi River. Chengdu is the
largest city in southwest China, located at 30°05′–31°26′
north latitude and 102°54′–104°53′ east longitude (Fig. 1).
Chengdu has a typical subtropical monsoonal humid and
mild climate with four distinct seasons. Chengdu is an
important city and economic center in western China.
The resident population is over 20 million [14].
Data collection
Daily non-accidental death counts from January 1, 2016,
to December 31, 2020, were obtained from the Popula-
tion Death Information Registration and Management
System (PDIRMS). Mortality data were obtained that
covered all mortality-related information for residents of
Chengdu city. In the PDIRMS, the death of a resident is
confirmed by a hospital or doctor at the resident’s home,
and mortality data are subsequently recorded in the
system. The causes of death are coded according to the
International Classification of Disease, Tenth Revision
(ICD-10). The mortality data are classified as death from
all causes (A00-R99), respiratory diseases (J00-J99), CVD
(I00-I99), cerebrovascular diseases (I60-I69), or IHD
(I20-25).
Data on airborne pollutants (particulate matter with
diameter 2.5 microns or less [­PM2.5], particulate matter
with diameter 10 microns or less [­PM10], sulfur dioxide
­[SO2], nitrogen dioxide [­NO2], and the daily 8-h mean
concentrations of ozone [­O3]) and daily meteorological
data (daily maximum, mean, and minimum temperatures
[°C] and mean relative humidity [RH%]) were derived
from 23 municipal environmental monitoring sites that
operated continuously from January 1, 2016, to Decem-
ber 31, 2020, in Chengdu.
Data analysis
We studied the association between mean ambient
temperature and daily counts of total non-accidental
mortality, CVD mortality, respiratory mortality, cer-
ebrovascular mortality, and IHD mortality. Several
studies have reported that the relationship between
temperature and mortality is not linear, but is instead
J-, V-, or U-shaped [15–18]. The distribution of tem-
perature effects over days or weeks after exposure is
often dealt with by establishing distributed-lag models
[19]. Thus, we used distributed-lag non-linear models
(DLNM) to estimate the non-linear and lag effects of
Xia et al.
BMC Public Health (2023) 23:149
Fig. 1  Location of study area in Chengdu city, southwestern China
temperature on daily mortality [15, 20, 21]. Long-term
trends, relative humidity and ambient air pollutants
­(PM2.5, ­O3) were controlled in the model as potential
confounders. A Poisson regression with a quasi-Poisson
function for the daily counts of deaths was constructed,
which was specified as follows:
Log[E(Yt)] = 𝛼 + cb(Tempt, l) + ns(PM2.5t, l, 3) + ns(O3t, l, 3) + ns(RelativeHumidityt, l, 3) + ns(Time, 8 ∗ 5) + DOWt,
where Yt is the observed number of daily deaths at
day t; α is the intercept; cb(Tempt,l) is the cross-basis
matrix produced by DLNM to model the non-linear
and distributed lag effects of ambient temperature from
the current day (lag0) to the fourteenth day (lag14);
ns(PM2.5t,l,3) is a natural cubic spline of ­ PM2.5 with
three degrees of freedom (df); ns(O3t,l,3) is a natural
cubic spline of O ­ 3 with three df; ns(RHt,l,3) is a natu-
ral cubic spline of relative humidity with three df; and
ns(Time,8*5) represents natural spline function with
eight df per year to control long-term and seasonal
trends [22]. DOWt is used to control the effect of day of
week. We calculated the relative risk (RR) of mortality
by comparing extreme cold (1st percentile of tempera-
ture) and extreme hot (99th percentile of temperature)
temperatures with the minimum mortality tempera-
ture. Akaike’s Information Criterion for quasi-Poisson
Page 3 of 12
models was used to selected the temperature threshold
and df for temperature and lag [23]. Additionally, we
stratified the analysis by gender, age group, education
level, and marital status through the above steps.
We tested the statistical significance of differences
between effect estimates of the strata of a potential effect
modifier (e.g., the difference between low education and
high education) by calculating the 95% confidence inter-
val (CI). The equation was as follows:

   2  2
 
Q1 − Q2 ± 1.96 SE1 + S E 2
where Q 1 and
2 Q2 are
 the2estimates for the two catego-
ries, and S E1 and S E
 2 are their respective standard
errors [24].
We used sensitive analyses to assess stability of the model
by changing the df of the model, including for long-term
trends (df = 6–9), major air pollutants, and relative humid-
ity (df = 3–5). All statistical tests were two-sided, and values
of p < 0.05 were considered statistically significant. All anal-
yses were performed using R software version 4.1.2, and the
“dlnm” package was used to create the DLNM.
Xia et al. BMC Public Health (2023) 23:149 Page 4 of 12

Results and ­SO2, and between mean temperature and relative

During the study period (January 1, 2016, to December
31, 2020), there were 343,313 non-accidental deaths in
total; 116,661 individuals (33.9%) died from CVD; 98,762
(28.8%) died from respiratory disease; 52,533 (15.3%)
died from cerebrovascular disease, and 27,464 individu-
als (8.0%) died from IHD. The daily average number of
deaths due to total non-accidental, CVD, respiratory,
cerebrovascular and IHD mortality was 188, 64, 54, 29,
and 15, respectively. Daily mean temperature and rela-
tive humidity were 16.8 °C (range: − 1.6 °C–30.1 °C) and
79.9% (range: 36.0%–99.0%), respectively. Daily mean
concentrations of air pollutants are shown in Table 1.
Table 2 shows the Spearman’s correlation coefficients of
air pollutants and meteorological conditions. The results
showed correlations between air pollutants and mete-
orological conditions, except between mean temperature
humidity.
The three-dimensional plots revealed higher relative
risks at hot and cold temperatures after fortnight lag.
Specifically, a non-linear relationship was found between
mean temperature and total non-accidental, CVD, respir-
atory, cerebrovascular and IHD deaths (Fig. 2 A–E).
Figure  3 shows the estimated cumulative effects of
mean temperature on total non-accidental, CVD, res-
piratory, cerebrovascular, and IHD mortality at lags of
0–14  days. The relationships of mean temperature with
total non-accidental, CVD, respiratory, cerebrovascular,
and IHD mortality were non-linear, with higher relative
risks at cold and hot temperatures. The temperature with
the minimum mortality risks for total non-accidental,
CVD, respiratory, cerebrovascular, and IHD were 21.5 °C,
21 °C, 20 °C, 21 °C, and 20.5 °C, respectively.

Table 1 Summary statistics of daily cause-specific mortality, air pollutants, and weather conditions in Chengdu, 2016–2020
(1827 days)
Variables Mean SD Min 1st 20% 50% 90% 99% Max

Total (non-accidental) 187.9 39.3 106.0 124.0 156.0 180.0 246.0 303.0 340.0
CVD 63.9 16.1 30.0 35.0 50.0 62.0 86.0 108.0 119.0
Respiratory 54.1 19.0 20.0 27.0 39.0 49.0 79.0 119.0 136.0
Cerebrovascular 28.8 7.8 10.0 14.0 22.0 28.0 40.0 48.0 65.0
IHD 15.0 5.3 3.0 5.0 10.0 14.0 22.0 30.0 39.0
Air pollutants
  ­PM2.5 (µg/m3) 49.5 34.7 3.0 8.1 22.4 39.6 95.0 170.1 259.8
  ­PM10 (µg/m3) 78.5 49.9 5.5 15.3 38.7 66.6 143.1 247.0 358.1
  ­SO2 (µg/m3) 10.0 4.8 3.4 4.0 5.8 8.6 16.6 24.8 29.8
  ­NO2 (µg/m3) 38.0 13.4 7.0 13.8 26.3 36.2 56.5 73.4 91.1
  ­O3 (µg/m3) 95.1 49.4 10.3 17.8 51.7 84.4 166.3 219.4 278.0
Meteorological variables
  Mean temperature (°C) 16.8 7.4 -1.6 3.0 8.6 17.1 26.3 29.0 27.0
  Minimum temperature (°C) 13.5 7.2 -6.2 -1.4 6.1 14.3 23.0 24.7 37.2
  Maximum temperature (°C) 79.9 8.2 3.3 6.1 12.8 21.7 32.0 35.6 99.0
  Relative humidity (%) 9.8 36.0 52.0 72.0 81.0 92.0 97.0
*
Min minimum, Max maximum, SD standard deviation

Table 2  Spearman’s correlation coefficients of weather conditions and air pollutants in Chengdu, China (2016–2020)
Variables PM2.5 O3 SO2 NO2 Mean temperature Relative
humidity

PM2.5 1
O3  − 0.20* 1
SO2 0.61* 0.16* 1
NO2 0.80*  − 0.17* 0.65* 1
Mean temperature  − 0.51* 0.72*  − 0.04  − 0.41* 1
Relative humidity  − 0.13*  − 0.49*  − 0.21*  − 0.16*  − 0.04 1
*
p < 0.05 for all correlation coefficients
 Xia et al. BMC Public Health (2023) 23:149
Fig. 2  Relative risks of cause-specific mortality by mean temperature (°C) and lag in Chengdu, China, during 2016–2020. A–E
Figure  4 shows the estimated risks of daily mortal-
ity associated with extreme cold (3 °C) and extreme hot
temperatures (29  °C) at lags of 0–14  days. Significant
effects of extreme cold temperatures were observed after
0–2  days and persisted for 7–10  days. The most signifi-
cant and strongest effects of extreme hot temperatures
were observed on the current day and lasted for only
0–3 days.
We calculated the overall effects of mean temperature
on total non-accidental, CVD, cerebrovascular, respira-
tory, and IHD mortality for lags of 0–1, 0–3, 0–7, and
0–14 days (Table 3). The results showed that the relative
risks associated with extreme cold temperature (1st per-
centile of temperature, 3.0 °C) over lags 0–14 days were
1.32 (95% CI: 1.19, 1.46) for total mortality, 1.45 (95% CI:
1.24, 1.68) for CVD morality, 1.28 (95% CI: 1.09, 1.50) for
Page 5 of 12
respiratory morality, 1.36 (95% CI: 1.09, 1.70) for cerebro-
vascular mortality, and 1.26 (95% CI: 0.95, 1.68) for IHD
morality, respectively. The relative risks associated with
extreme hot temperature (99th percentile of temperature,
29 °C) over lags 0–3 days were 1.22 (95% CI: 1.17, 1.28)
for total mortality, 1.40 (95% CI: 1.30, 1.50) for CVD
morality, 1.34 (95% CI: 1.24, 1.46) for respiratory moral-
ity, 1.33 (95% CI: 1.20, 1.47) for cerebrovascular mortal-
ity, and 1.38 (95% CI: 1.20, 1.58) for IHD mortality.
Figure 5 presents the relative risks of low and high tem-
peratures for disease-related mortality in subgroups by
sex, age, education level, and marital status. The results
indicated that the relationship between temperature
and mortality was more pronounced among adults over
85 years of age, women, individuals with low education,
and those with other marital status.
Xia et al. BMC Public Health (2023) 23:149
Fig. 3  Estimated relative risks of mean temperature (°C) over lags 0–14 days on total non-accidental, CVD, respiratory, cerebrovascular, and IHD
mortality. Blue lines are mean relative risks and gray areas are 95% CIs of risk estimates
Discussion
In the present study, we examined the effects of ambi-
ent temperature on cause-specific mortality among
older adults in Chengdu, China, during the years 2016–
2020. The results showed that both cold and high tem-
peratures increased the risk of mortality. J-shaped and
U-shaped relationships were found between tempera-
ture and all etiology-specific mortality categories. The
relationship between temperature and mortality was
more pronounced among adults over 85  years of age,
women, individuals with low education, and those who
were divorced, which is consistent with previous stud-
ies [19, 25, 26]. Some prior research reported a signifi-
cantly stronger association between temperature and
disease-related mortality in women than in men [27, 28],
while some reported the opposite [29]. Furthermore, our
results showed that the effects of cold temperatures on
mortality were much larger than those of hot tempera-
tures, with the effects of cold temperatures occurring
Page 6 of 12
after 0–2 days and lasting 7–10 days, whereas the effects
of high temperatures were short-lived. Similar results
have been found in previous studies in Beijing [30],
Guangzhou [31], Hong Kong [32], and Thailand [33].
Many studies have confirmed that high and low tem-
peratures lead to an increased risk of death, that people
in different regions are adapted to different climates, and
that the degree of influence of temperature varies accord-
ing to geography, population, and environmental con-
ditions [8, 16, 34]. Our study confirmed the findings of
other studies. It is worth noting that in comparing the
99th percentile of temperature (29  °C) to the minimum
mortality temperature (21.5  °C), we found that expo-
sure to 2-day average high temperatures increased non-
accidental overall mortality by 18% (95% CI: 13%–22%),
which is greater than that reported in a study conducted
in Jinan (3.8% 95% CI: 2.6%–5.0%) [35]. However, a
study in Wuhan found a similar heat effect, with a 17.7%
(12.6%–22.9%) increase in total non-accidental mortality
 Xia et al. BMC Public Health (2023) 23:149 Page 7 of 12

Fig. 4  Estimated relative risks of extreme cold and extreme hot temperatures over lags 0–14 days on total non-accidental, CVD, respiratory,
cerebrovascular, and IHD mortality. Extreme cold temperature: 1st percentile of temperature, 3.0 °C; extreme hot temperature: 99th percentile of
temperature, 29.0 °C

at 99% temperature (34.7  °C) relative to the minimum
mortality temperature (31.7  °C) [36]. Consistent with
the results of the study in Wuhan [30], we did not find
a significant effect of low temperature on IHD mortality,
possibly because the relatively small number of deaths
owing to IHD limited our ability to detect an association
between temperature and IHD.
Many previous studies have demonstrated that older
adults are more sensitive to the effects of temperature
[35, 37, 38]. Our research indicated that older adults are
more susceptible to the effects of cold temperatures. This
is consistent with the findings of a study in Kuwait [39],
where the incidence of death owing to CVD was higher in
old people, especially during cold periods. Several studies
have reported that sensitivity to heat and cold exposure
may vary among populations in different regions owing
to adaptation to local climate [5, 8, 40]. Population age
structure, socioeconomic conditions, education, health
care, infrastructure development, housing quality, and
air conditioning use also influence the effects of hot and
cold temperatures on human health [8, 41, 42]. For exam-
ple, an Australian study found a diminished association
Xia et al. BMC Public Health (2023) 23:149
Table 3  Cumulative relative risks of extreme hot and extreme cold temperatures on daily mortality, compared with the minimum
mortality temperature
Lag Total mortality Cardiovascular mortality
Extreme hot 0–1 1.18 (1.13,1.22)* 1.31 (1.24,1.39)*
0–3 1.22 (1.17,1.28)* 1.40 (1.30,1.50)*
0–7 1.17 (1.11,1.23)* 1.30 (1.20,1.41)*
0–14 1.14 (1.06,1.23)* 1.28 (1.14,1.43)*
Extreme cold 0–1 0.94 (0.90,0.99) 1.01 (0.94,1.08)
0–3 1.00 (0.94,1.06) 1.06 (0.97,1.15)
0–7 1.18 (1.10,1.26)* 1.23 (1.11,1.36)*
0–14 1.32 (1.19,1.46)* 1.45 (1.24,1.68)*
Extreme cold temperature: 1st percentile of temperature, 3 °C; extreme hot temperature: 99th percentile of temperature, 29 °C; minimum mortality temperature for
total non-accidental, cardiovascular, respiratory, cerebrovascular, and ischemic heart disease mortality: 21.5 °C, 21 °C, 20 °C, 21 °C, and 20.5 °C, respectively
*
p < 0.05
between temperature and cardiovascular mortality
in people over the age of 85  years who lived in areas of
higher socioeconomic status [43].
When analyzing the effect of temperature on mortal-
ity according to cause of death, we found that the effects
of hot and cold temperatures on CVD-related mortality
were all greater than those of non-accidental death and
mortality related to respiratory diseases, cerebrovascular
diseases, and IHD, which is consistent with the results
of a study in Suzhou, China [19]. CVD was the lead-
ing cause of death among elderly residents of Chengdu,
accounting for 33.9% of all registered deaths during the
study period. In addition, previous studies have con-
firmed that the effect of temperature on mortality owing
to different diseases is not consistent. For example, many
studies have found that the effects of heat and cold was
more pronounced in CVD mortality. However, in a study
in Jinan, China, low temperature was found to signifi-
cantly increase mortality from respiratory diseases; at
high temperatures, there was an increase in deaths from
respiratory diseases, but the effect was not statistically
significant [44]. Additionally, a study in Chiang Mai,
Thailand, showed that high temperatures significantly
increased mortality from respiratory diseases [45].
Our findings suggested that people over 85 years of age
are more sensitive to temperature and have lower mor-
tality rates in the age group 65–74  years, showing more
sensitivity to temperature with older age, which confirms
previous findings by Breitner [25]. As mentioned in the
published literature, concomitant diseases and weakened
thermoregulatory mechanisms among older adults may
be important factors contributing to mortality, and older
people with chronic diseases, such as cardiovascular or
respiratory diseases, are usually more vulnerable to hot
weather than other groups [46, 47]. Another important
finding was that the cold effect was greater than the heat
Page 8 of 12
Respiratory mortality Cerebrovascular mortality IHD mortality
1.25 (1.17,1.34)* 1.26 (1.16,1.38)* 1.28 (1.14,1.44)*
1.34 (1.24,1.46)* 1.33 (1.20,1.47)* 1.38 (1.20,1.58)*
1.30 (1.19,1.43)* 1.26 (1.12,1.41)* 1.29 (1.10,1.15)*
1.27 (1.11,1.45)* 1.25 (1.06,1.47)* 1.24 (0.98,1.56)
0.91 (0.84,0.97) 1.07 (0.96,1.18) 0.93 (0.82,1.06)
0.96 (0.88,1.05) 1.11 (0.99,1.26) 0.97 (0.83,1.13)
1.16 (1.04,1.29)* 1.23 (1.06,1.43)* 1.09 (0.90,1.31)
1.28 (1.09,1.50)* 1.36 (1.09,1.70)* 1.26 (0.95,1.68)
effect for all disease mortality in those aged 85 years and
older, except for the cold effect on respiratory disease
mortality, which was slightly smaller than the heat effect.
In China, the number of people aged 65 years and older
reached 190 million in 2020, accounting for 13.5% of the
total population. It is expected that by 2050, the propor-
tion of people aged 65 years and above will exceed 25%.
China’s disabled elderly population will also exceed 97.5
million by approximately 2050. For disabled older peo-
ple, the inability to take effective measures when they are
in hot or cold conditions can lead to an increased risk of
death. Considering that older people are more sensitive
to low temperatures and given the rapid aging of Chi-
na’s population, the adverse effects of low temperatures
on older adults require attention. In addition, tailored
measures of treatment care can substantially improve the
health and wellbeing of older populations.
We found significantly greater heat effects in women
than in men, and greater cold effects in men than in
women, although the differences were not significant.
This is similar to the results of a previous study con-
ducted in Barcelona, Spain [26]. There are also some dif-
ferences in previous studies regarding correction owing
to the effect of sex on mortality. Some studies have shown
a significantly stronger association between tempera-
ture and disease-related mortality in women than in men
[27, 28] whereas others have found a greater risk in men
[29]. In contrast, several other studies have found no sex
differences; differences in temperature effects between
women and men were influenced by the study site and
study population [19, 48]. For example, the reported
risk of heat effects on total mortality was greater among
men in Shanghai, China, and among women of all ages in
Fukuoka, Japan [49, 50].
Another conclusion that can be drawn from our results
is that older Chengdu residents with low educational
 Xia et al. BMC Public Health (2023) 23:149 Page 9 of 12

Fig. 5  Relative risk (RR) of cold and heat on mortality associated with different diseases among older adults in Chengdu, China. A–E) total
non-accidental, cardiovascular, respiratory, cerebrovascular, and ischemic heart disease-related deaths. Low education level: illiterate and primary;
high education level: junior high school and above. Other marital status: widowed, divorced, and never married

attainment are more susceptible to temperature-related
mortality, with both hot and cold effects being signifi-
cant. Past studies have also confirmed that people with
lower education levels are more susceptible to tempera-
ture-related mortality. For example, Yang et al. [51] found
that the effect of heat and cold was greater for those
with lower levels of education. Li et  al. [35] reported
that low levels of education were associated with heat-
related mortality. Wang et al. [19] did not observe a sig-
nificant modifying effect of education level, although
they reported a greater vulnerability of residents with
low levels of education to temperature-related mortality.
One explanation for this is that people with low levels of
education may be more exposed to cold or heat and may
have poor living and housing conditions, limited access
to health care, and may lack knowledge about precau-
tions against cold and heat exposure, which may contrib-
ute to their increased risk of death [52, 53].
Another important finding is that marital status can
also explain the association between temperature and
Xia et al. BMC Public Health (2023) 23:149
mortality. We divided marital status into married and
other (widowed, divorced, and never married). We found
that non-married people had higher mortality associated
with heat and cold than married people, although the dif-
ference in the cold effect was not significant. This is con-
sistent with results of a study conducted in São Paulo,
Brazil, showing that widowed individuals had higher
mortality associated with cold and heat [54]. Many stud-
ies have reported that marriage has a "protective effect"
on health [55–58]. We speculate that married people are
better off financially and can share health benefits with
their spouse; married couples may also be more emotion-
ally supportive of each other and have a more positive
and optimistic attitude toward life.
In a hot environment, increased body temperature
redistributes blood flow to the skin, and the body releases
heat through mechanisms such as skin sweat secretion
and vasodilation, which can lead to a loss of salt and
water from the body. If not adequately replenished, this
can lead to dehydration, which decreases blood volume
and can ultimately increase cardiovascular strain [7]. Past
evidence suggests that the longer a heat wave lasts, the
greater the risk of CVD death. Older adults, women, and
outdoor workers have higher rates of death [59]. Notably,
when the body’s thermoregulatory capacity is diminished
and the internal temperature is too high (39  °C–40  °C),
kidney and liver damage, as well as central nervous sys-
tem damage, can occur [7].
In contrast, when in a cold environment, skin tempera-
ture decreases and vascular tone decreases, causing the
body to increase metabolic heat production and begin to
shiver to maintain the body’s core temperature [60]. How-
ever, it has been shown that an inability to maintain core
temperature in severely cold environments among people
over 65 years of age can lead to tissue damage [61].
The robustness of our model was tested using sensi-
tivity analysis. For temporal trends in the model from 6
to 9, airborne pollutants and relative humidity from 3 to
5 were included (Tables S1 and S2). The RRs calculated
according to different df were similar. Therefore, the
results calculated by the model are reliable.
To our knowledge, this was the first study to explore
the association of multiple indicators for temperature
with mortality from four specific diseases in the central
Sichuan basin of southwest China. Based on the advan-
tages of the PDIRMS, we used realistic and reliable
mortality data for the entire city. With current global
warming and the frequent occurrence of extreme weather
events, the present findings may be critical in reducing
temperature-related mortality and provide a theoreti-
cal basis for follow-up research. However, this study also
has some limitations. First, the study was conducted in
Page 10 of 12
the central Sichuan basin, and the data were from only
one city, which makes it difficult to generalize our results
to other regions considering geographic characteristics,
meteorological conditions, and other uncontrollable
factors. Second, data for cause of death were classified
according to the ICD-10 code on the death certificate,
which may be biased. Third, pollutant and meteorologi-
cal data were from fixed municipal environmental moni-
toring sites rather than individual exposures, which may
result in some unavoidable assessment errors. Further
research for evaluating the overall situation in multiple
cities is needed to more accurately analyze the effects
of ambient temperature on mortality among elderly
residents. Additionally, the results would be more reli-
able if the confounding of mortality risk factors could be
excluded in future related studies.
Conclusions
The results of this study showed that exposure to hot
and cold temperatures in Chengdu was associated with
increased mortality, with people over 85  years old,
women, those with low education levels, and unmar-
ried individuals being more affected by hot and cold
temperatures.
Abbreviations
PM2.5 Particulate matter < 2.5 μm in aerodynamic diameter
SO2 Sulfur dioxide
NO2 Nitrogen dioxide
CO Carbon monoxide
O3-8 h Daily 8-h mean concentrations of ozone
RH Relative humidity
MMT Minimum mortality temperature
DLNM Distributed-lag non-linear models
CVD Cardiovascular disease
IHD Ischemic heart disease
SD Standard deviation
RR Relative risk
CI Confidence interval
Supplementary Information
The online version contains supplementary material available at https://​doi.​
org/​10.​1186/​s12889-​022-​14931-x.
Additional file 1: Table S1. Time trend freedom 6-9 Table S2. Air pollut-
ants and relative humidity trend freedom 3-5.
Acknowledgements
We thank the Chengdu Center for Disease Control and Prevention for provid-
ing data.
Authors’ contributions
XYZ and SCL coordinated the study, performed data analysis, and drafted
the manuscript;LY, JXY, RSJ and CY contributed to the statistical analyses;
GXF,HW,LMJ,XR,WXY and PXJ assisted in obtaining data; CJY coordinated
the study and edited the manuscript;and LZ organized and coordinated the
study and edited the manuscript. All authors have read and approved the final
manuscript.
 Xia et al. BMC Public Health (2023) 23:149
Funding
This study was funded by the Sichuan Provincial Cadre Health Care Research
Project (No. 2021–1801), and the Sichuan Provincial Cadre Health Care
Research Project (No. ZH2018-1801).
Availability of data and materials
The datasets generated and/or analysed during the current study are not
publicly available due involvement of detailed information on regional popu-
lation deaths, which needs to be kept confidential but are available from the
corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
All procedures performed in studies involving human participants were in
accordance with the ethical standards of the institutional and/or national
research committee and with the 1964 Helsinki. We declare the informed
consent is obtained from all the participants. This study was approved by
the Ethics Committee of Sichuan Provincial Center for Disease Control and
Prevention.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Author details
1
 Sichuan Provincial Center for Disease Control and Prevention, No.6,
Zhongxue Road, Wuhou District, Chengdu 610041, China. 2 School of Pub-
lic Health, Chengdu Medical College, No.783, Xindu Road, Xindu District,
Chengdu 610500, China. 3 Chengdu Center for Disease Control and Pre-
vention, No.6, Longxiang Road, Wuhou District, Chengdu 610041, China.
4
 Zigong Center for Disease Control and Prevention, No.826, Huichuan Road,
Ziliujing District, Zigong 643000, China. 5 Panzhi Hua Center for Disease
Control and Prevention, Dong District, No.996, Jichang Road617067, Panzhi
Hua, China. 6 Guangyuan Center for Disease Control and Prevention, No.996,
Binhebei RoadLizhou District, Guangyuan 628017, China. 7 Mianyang Center
for Disease Control and Prevention, Gaoxin District, No.50, Mianxingdong
Road, Mianyang 621000, China. 8 Yaan Center for Disease Control and Preven-
tion, No.9, Fangcao Road, Yucheng District, Yaan 625000, China.
Received: 3 August 2022 Accepted: 22 December 2022
References
1. Guo Y, Barnett AG, Pan X, Yu W, Tong S. The impact of temperature on
mortality in Tianjin, China: a case-crossover design with a distributed lag
nonlinear model. Environ Health Perspect. 2011;119(12):1719–25.
2. Kouis P, Kakkoura M, Ziogas K, Paschalidou AΚ, Papatheodorou SI. The
effect of ambient air temperature on cardiovascular and respiratory
mortality in Thessaloniki. Greece Sci Total Environ. 2019;647:1351–8.
3. Lee H, Myung W, Kim H, Lee E-M, Kim H. Association between ambient
temperature and injury by intentions and mechanisms: A case-cross-
over design with a distributed lag nonlinear model. Sci Total Environ.
2020;746:141261.
4. Change, Climate I: "The physical science basis." Contribution of working
group I to the fifth assessment report of the intergovernmental panel on
climate change. 2013.
5. Ha J, Shin Y, Kim H. Distributed lag effects in the relationship between
temperature and mortality in three major cities in South Korea. Sci Total
Environ. 2011;409(18):3274–80.
6. Rodrigues M, Santana P, Rocha A. Effects of extreme temperatures on
cerebrovascular mortality in Lisbon: a distributed lag non-linear model.
Int J Biometeorol. 2019;63(4):549–59.
7. Kristie L, Ebi AC. Peter Berry, Carolyn Broderick, Richard de Dear,
George Havenith, Yasushi Honda, R Sari Kovats, Wei Ma, Arunima
Malik, Nathan B Morris, Lars Nybo, Sonia I Seneviratne, Jennifer Vanos,
Page 11 of 12
Ollie Jay: Hot weather and heat extremes health risks. The Lancet.
2021;398(10301):698–708.
8. Silveira IH, Oliveira BFA, Cortes TR, Junger WL. The effect of ambient
temperature on cardiovascular mortality in 27 Brazilian cities. Sci Total
Environ. 2019;691:996–1004.
9. Aklilu D, Wang T, Amsalu E, Feng W, Li Z, Li X, Tao L, Luo Y, Guo M, Liu
X, et al. Short-term effects of extreme temperatures on cause specific
cardiovascular admissions in Beijing. China Environ Res. 2020;186:109455.
10. Cai W, Zhang C, Zhang S, Bai Y, Callaghan M, Chang N, Chen B, Chen H,
Cheng L, Cui X, et al. The 2022 China report of the Lancet Countdown on
health and climate change: leveraging climate actions for healthy ageing.
Lancet Public Health. 2022;7(12):e1073–90.
11. Zhaoxing Tian SL. JZ, JJKJaYG: Ambient temperature and coronary heart
disease mortality in Beijing, China a time series study. Environ Health
Perspect. 2012;11(1):1–7.
12. Yang C, Meng X, Chen R, Cai J, Zhao Z, Wan Y, Kan H. Long-term variations
in the association between ambient temperature and daily cardiovascu-
lar mortality in Shanghai. China Sci Total Environ. 2015;538:524–30.
13. Lian T, Fu Y, Sun M, Yin M, Zhang Y, Huang L, Huang J, Xu Z, Mao C, Ni J, et al.
Effect of temperature on accidental human mortality: A time-series analysis
in Shenzhen, Guangdong Province in China. Sci Rep. 2020;10(1):8410.
14. Resident Population (Year-end) by region. Sichuan Statistical Yearbook
[http://​tjj.​sc.​gov.​cn/​scstjj/​c1058​55/​nj.​shtml]
15. Gasparrini A, Armstrong B, Kenward MG. Distributed lag non-linear mod-
els. Stat Med. 2010;29(21):2224–34.
16. Chung JY, Honda Y, Hong YC, Pan XC, Guo YL, Kim H. Ambient tempera-
ture and mortality: an international study in four capital cities of East Asia.
Sci Total Environ. 2009;408(2):390–6.
17. McMichael AJ, Wilkinson P, Kovats RS, Pattenden S, Hajat S, Armstrong B,
Vajanapoom N, Niciu EM, Mahomed H, Kingkeow C, et al. International
study of temperature, heat and urban mortality: the “ISOTHURM” project.
Int J Epidemiol. 2008;37(5):1121–31.
18. Medina-Ramon M, Schwartz J. Temperature, temperature extremes, and
mortality: a study of acclimatisation and effect modification in 50 US
cities. Occup Environ Med. 2007;64(12):827–33.
19. Wang C, Chen R, Kuang X, Duan X, Kan H. Temperature and daily
mortality in Suzhou, China: a time series analysis. Sci Total Environ.
2014;466–467:985–90.
20. Armstrong B. Models for the relationship between ambient temperature
and daily mortality. Epidemiology. 2006;17(6):624–31.
21. Gasparrini A. Distributed lag linear and non-linear models in R the pack-
age dlnm. J Stat Softw. 2011;43(8):1.
22. Achebak H, Devolder D, Ballester J. Trends in temperature-related
age-specific and sex-specific mortality from cardiovascular diseases
in Spain: a national time-series analysis. The Lancet Planetary Health.
2019;3(7):e297–306.
23. Roger D. Peng FDaTAL: Model choice in time series studies of air pollution
and mortality. J R Stat Soc A Stat Soc. 2006;169(2):179–203.
24. GENTLEMAN NSaJF: On judging the significance of differences by
examining the overlap between confidence intervals. Am Stat.
2001;55(3):182–6.
25. Breitner S, Wolf K, Devlin RB, Diaz-Sanchez D, Peters A, Schneider A. Short-
term effects of air temperature on mortality and effect modification by
air pollution in three cities of Bavaria, Germany: a time-series analysis. Sci
Total Environ. 2014;485–486:49–61.
26. Mari-Dell’Olmo M, Tobias A, Gomez-Gutierrez A, Rodriguez-Sanz M. Garcia
de Olalla P, Camprubi E, Gasparrini A, Borrell C: Social inequalities in the
association between temperature and mortality in a South European
context. Int J Public Health. 2019;64(1):27–37.
27. Deng J, Hu X, Xiao C, Xu S, Gao X, Ma Y, Yang J, Wu M, Liu X, Ni J, et al.
Ambient temperature and non-accidental mortality: a time series study.
Environ Sci Pollut Res Int. 2020;27(4):4190–6.
28. Seposo XT, Dang TN, Honda Y. Evaluating the Effects of Temperature on
Mortality in Manila City (Philippines) from 2006–2010 Using a Distributed
Lag Nonlinear Model. Int J Environ Res Public Health. 2015;12(6):6842–57.
29. Son JY, Lee JT, Anderson GB, Bell ML. Vulnerability to temperature-related
mortality in Seoul, Korea. Environ Res Lett. 2011;6(3):034027.
30. Zhang Y, Li S, Pan X, Tong S, Jaakkola J, Gasparrini A, Guo Y. S W: The
effects of ambient temperature on cerebrovascular mortality an epide-
miologic study in four climatic zones in China. Environ Health Perspect.
2014;13(1):1–12.
Xia et al. BMC Public Health (2023) 23:149 Page 12 of 12

31. Wu W, Xiao Y, Li G, Zeng W, Lin H, Rutherford S, Xu Y, Luo Y, Xu X, Chu C, 54. Son JY, Gouveia N, Bravo MA, de Freitas CU, Bell ML. The impact of tem-
et al. Temperature-mortality relationship in four subtropical Chinese cit- perature on mortality in a subtropical city: effects of cold, heat, and heat
ies: a time-series study using a distributed lag non-linear model. Sci Total waves in Sao Paulo. Brazil Int J Biometeorol. 2016;60(1):113–21.
Environ. 2013;449:355–62. 55. Robards J, Evandrou M, Falkingham J, Vlachantoni A. Marital status, health
32. Yi W, Chan AP. Effects of temperature on mortality in Hong Kong: a time and mortality. Maturitas. 2012;73(4):295–9.
series analysis. Int J Biometeorol. 2015;59(7):927–36. 56. Goldman YHaN. Mortality differentials by marital status an international
33. Denpetkul T, Phosri A. Daily ambient temperature and mortality in comparison. Demography. 1990;27(2):233–50.
Thailand: Estimated effects, attributable risks, and effect modifications by 57. NOREEN GOLDMAN SKaRW. Marital status and health among the elderly.
greenness. Sci Total Environ. 2021;791:148373. Soc Sci Med. 1995;40(12):1717–30.
34. Shindell D, Zhang Y, Scott M, Ru M, Stark K, Ebi KL. The Effects of Heat 58. Eaker ED, Sullivan LM, Kelly-Hayes M, D’Agostino RB Sr, Benjamin EJ. Mari-
Exposure on Human Mortality Throughout the United States. Geohealth. tal status, marital strain, and risk of coronary heart disease or total mortal-
2020;4(4):e2019GH000234. ity: the Framingham Offspring Study. Psychosom Med. 2007;69(6):509–13.
35. Li J, Xu X, Yang J, Liu Z, Xu L, Gao J, Liu X, Wu H, Wang J, Yu J, et al. Ambi- 59. Yin Q, Wang J. The association between consecutive days’ heat wave and
ent high temperature and mortality in Jinan, China: A study of heat cardiovascular disease mortality in Beijing, China. BMC Public Health.
thresholds and vulnerable populations. Environ Res. 2017;156:657–64. 2017;17(1):223.
36. Zhang Y, Li C, Feng R, Zhu Y, Wu K, Tan X, Ma L. The Short-Term Effect 60. Karpov VY, Zavalishina SY, Bakulina ED, Dorontsev AV, Gusev AV, Fedorova
of Ambient Temperature on Mortality in Wuhan, China: A Time-Series TY, Okolelova VA. The Physiological Response of the Body to Low Tem-
Study Using a Distributed Lag Non-Linear Model. Int J Environ Res Public peratures. J Biochem Technol. 2021;12(1):27–31.
Health. 2016;13(7):722. 61. Degroot DW, Kenney WL. Impaired defense of core temperature in aged
37. Basu R, Ostro BD. A multicounty analysis identifying the populations humans during mild cold stress. Am J Physiol Regul Integr Comp Physiol.
vulnerable to mortality associated with high ambient temperature in 2007;292(1):R103-108.
California. Am J Epidemiol. 2008;168(6):632–7.
38. Lin YK, Ho TJ, Wang YC. Mortality risk associated with temperature and
prolonged temperature extremes in elderly populations in Taiwan. Envi- Publisher’s Note
ron Res. 2011;111(8):1156–63. Springer Nature remains neutral with regard to jurisdictional claims in pub-
39. Alahmad B, Shakarchi AF, Khraishah H, Alseaidan M, Gasana J, Al-Hemoud lished maps and institutional affiliations.
A, Koutrakis P, Fox MA. Extreme temperatures and mortality in Kuwait:
Who is vulnerable? Sci Total Environ. 2020;732:139289.
40. Iniguez C, Ballester F, Ferrandiz J, Perez-Hoyos S, Saez M, Lopez A, Tempro
E. Relation between temperature and mortality in thirteen Spanish cities.
Int J Environ Res Public Health. 2010;7(8):3196–210.
41. Kenny GP, Yardley J, Brown C, Sigal RJ, Jay O. Heat stress in older
individuals and patients with common chronic diseases. CMAJ.
2010;182(10):1053–60.
42. Wang C, Zhang Z, Zhou M, Zhang L, Yin P, Ye W, Chen Y. Nonlinear
relationship between extreme temperature and mortality in different
temperature zones: A systematic study of 122 communities across the
mainland of China. Sci Total Environ. 2017;586:96–106.
43. Lu P, Zhao Q, Xia G, Xu R, Hanna L, Jiang J, Li S, Guo Y: Temporal trends
of the association between ambient temperature and cardiovascular
mortality: a 17-year case-crossover study. Environ Res Lett. 2021;16(4).
44. Han J, Liu S, Zhang J, Zhou L, Fang Q, Zhang J, Zhang Y. The impact of
temperature extremes on mortality: a time-series study in Jinan, China.
BMJ Open. 2017;7(4):e014741.
45. Guo Y. KPaST: Effects of temperature on mortality in Chiang Mai city,
Thailand a time series study. Environ Health. 2012;11(1):1–9.
46. Fouillet A, Rey G, Laurent F, Pavillon G, Bellec S, Guihenneuc-Jouyaux C,
Clavel J, Jougla E, Hemon D. Excess mortality related to the August 2003
heat wave in France. Int Arch Occup Environ Health. 2006;80(1):16–24.
47. Vaneckova P, Beggs PJ, Jacobson CR. Spatial analysis of heat-related
mortality among the elderly between 1993 and 2004 in Sydney. Australia
Soc Sci Med. 2010;70(2):293–304.
48. Ma W, Yang C, Tan J, Song W, Chen B, Kan H. Modifiers of the tem-
perature-mortality association in Shanghai. China Int J Biometeorol.
2012;56(1):205–7.
49. Huang W, Kan H, Kovats S. The impact of the 2003 heat wave on mortality
in Shanghai. China Sci Total Environ. 2010;408(11):2418–20.
Ready to submit your research ? Choose BMC and benefit from:
50. Onozuka D, Hagihara A. Variation in vulnerability to extreme-tempera-
ture-related mortality in Japan: A 40-year time-series analysis. Environ Res.
• fast, convenient online submission
2015;140:177–84.
51. Yang J. C-QO, Yan Ding, Y-XZaP-YC: Daily temperature and mortality • thorough peer review by experienced researchers in your field
a study of distributed lag non-linear effect and effect modification in • rapid publication on acceptance
Guangzhou. Environ Health. 2012;11(1):1–9.
• support for research data, including large and complex data types
52. Medina-Ramon M, Zanobetti A, Cavanagh DP, Schwartz J. Extreme
temperatures and mortality: assessing effect modification by personal • gold Open Access which fosters wider collaboration and increased citations
characteristics and specific cause of death in a multi-city case-only analy- • maximum visibility for your research: over 100M website views per year
sis. Environ Health Perspect. 2006;114(9):1331–6.
53. Mirabelli MAMaM: The potential impacts of climate variability and change At BMC, research is always in progress.
on temperature-related morbidity and mortality in the United States.
Environ Health Perspect. 2001;109(suppl 2):185–9. Learn more biomedcentral.com/submissions